To evaluate the usefulness of CT scanning before carotid endarterectomy, a prospective study was performed on 469 consecutive patients considered for carotid endarterectomy during a 5-year period. All patients underwent carotid duplex scanning and CT scanning before carotid arteriography. Two hundred thirty-seven patients (51%) had transient ischemic attacks, 109 (23%) had a prior stroke, and 122 (26%) were asymptomatic. Results of the CT scan were abnormal in 68 (62%) of the 109 patients with stroke. Fifty-one of the 360 patients (14%) without a clinical history of stroke had an abnormal CT scan outcome. Of patients with a stroke documented by CT scanning, 27 had lacunar infarcts, and 92 had cortical infarcts; these findings did not change surgical management in any patient. CT scanning did not reveal any unsuspected infarcts or tumors. Two hundred thirty carotid endarterectomies were performed on 206 patients. Forty-seven patients (23%) in the operative group had abnormal CT scan findings, but the scan did not influence operative decisions or timing in any case. Seventy-two patients (27%) in the nonoperative group had abnormal CT scan results, but CT scan findings did not exclude any patient from arteriography or surgery. Three perioperative strokes (1.3%) occurred. CT scan findings did not correlate with postoperative neurologic complications. Cost of CT scanning was one-half million dollars in our study alone. Routine CT scanning is unnecessary before carotid endarterectomy and is not cost-effective.